The `Walking' Megalithic Statues (Moai) of Easter Island

Journal of Archaeological Science 40 (2013) 2859e2866

Contents lists available at SciVerse ScienceDirect

Journal of Archaeological Science

journal homepage: http://www.elsevier.com/locate/jas

The ‘walking’ megalithic statues (moai) of Easter Island

Carl P. Lipo a,b,*, Terry L. Hunt c, Sergio Rapu Haoa d a IIRMES, California State University Long Beach, 1250 Bellﬂower Blvd., Long Beach, CA 90840, USA b Department of Anthropology, California State University Long Beach, 1250 Bellﬂower Blvd., Long Beach, CA 90840, USA c Department of Anthropology, University of Hawai’i, 2424 Maile Way, Saunders Hall 346, Honolulu, HI 96822-2223, USA d Instituto de Estudios Oceanicos, Hangaroa, Rapa Nui/Isla de Pascua, Chile article info abstract

Article history: Explaining how the monumental statues (moai) of Easter Island were transported has remained open to Received 25 July 2012 debate and speculation, including their resource expenditures and role in deforestation. Archaeological Received in revised form evidence including analysis of moai variability, particularly those abandoned along ancient roads, indi- 20 September 2012 cates transport was achieved in a vertical position. To test this proposition we constructed a precise Accepted 20 September 2012 three-dimensional 4.35 metric ton replica of an actual statue and demonstrate how positioning the center of mass allowed it to fall forward and rock from side to side causing it to ‘walk.’ Our experiments Keywords: reveal how the statue form was engineered for efﬁcient transport by a small number of individuals. Easter Island Rapa Nui Ó 2012 Elsevier Ltd. All rights reserved. Megaliths Moai Paciﬁc prehistory Eastern Polynesia

1. Introduction Understanding the role of carving and transporting hundreds of multi-ton moai remains central to the island’s prehistory. In this The ancient Polynesians of Easter Island carved nearly 1000 paper, we draw on detailed analysis of ancient statues (n 961 ¼ massive statues (moai) from volcanic hyalotuff bedrock and trans- total; see Torres Hochstetter et al., 2011), with a focus on those ported about 500 along roads traversing a rugged landscape to abandoned along roadways (Lipo and Hunt, 2005; n 62 statues; ¼ monumental sites around the island, some as far as 16e18 km see Table S1) during their transport. Based on our research, we distance (Fig. 1). The largest of these statues is over ca. 10 m tall, show that movement in a vertical ‘walking’ motion explains weighs approximately 74 metric tons and was moved over 5 km. systematic variations in moai found in the quarry, on ancient Early European visitors and later researchers have assumed moai roadways, and those found on platforms (ahu). To test a proposition carving and transport required a large labor force, a correspond- for vertical transport, we used a precisely scaled ﬁve-ton road- ingly large population, and a substantial expenditure of natural statue replica and show experimentally how relatively few people resources such as agricultural surplus and trees for rollers or other achieve movement in a ‘walking’ motion. Our new explanation of transport devices. The notion of an irrational statue cult fueled by statue transport accounts for systematic variability in the statues overpopulation and overexploitation of resources has formed and highlights the remarkable engineering skills of the prehistoric foundations of arguments for “ecocide” and seemed to provide Easter Islanders. a coherent model for Easter’s prehistory (Diamond, 2005, 2007). However, recent research and new evidence calls into question the 2. Previous investigations longstanding notions of “ecocide” and population collapse before European contact (Hunt and Lipo, 2011). The question of how the multi-ton moai of Easter were transported has puzzled visitors and researchers for centuries. No visitors to the island ever witnessed the process, leaving much to an array of speculations. The islanders’ oral traditions have long * Corresponding author. IIRMES, California State University Long Beach, 1250 recounted that the statues ‘walked’. Thomson (1889), for example, Bellﬂower Blvd., Long Beach, CA 90840, USA. Tel.: 1 562 985 2393; fax: 1 562 985 þ þ 4379. was told the statues were “endowed with power to walk about in E-mail addresses: [email protected], [email protected] (C.P. Lipo). the darkness.” Metraux (1940: 240) recounts, “the huge blocks

Fig. 1. Moai and moai roads of Easter Island. walked for a distance and then stopped.” However, oral tradition 3. Roads and statues offers no detailed explanation of how this ‘walking’ was achieved or to what degree it was meant metaphorically. Despite longstanding questions of how statues were moved, Modern experiments began with Heyerdahl’s efforts in the surprisingly little attention has been paid to the archaeological 1950s, including simply dragging them (Heyerdahl,1989). To resolve record directly related to transport or variability in the statues. In problems of friction and damage to statues, later efforts employed 2004, we undertook a study of prehistoric roads constructed as wooden sledges, pods, rollers, and sliders in various configurations. paths for the transport of moai (Lipo and Hunt, 2005, Fig. 1). First Among those proposing use of wooden devices, the debate focused noted in 1917 by Routledge (1919), these paths form linear features on whether statues were moved in a horizontal or vertical position approximately 4.5 m wide extending out from the quarry at Rano (Love, 1990, 2000; Van Tilburg and Ralston, 2005). In contrast, Pavel Raraku. Over 25 km of these roads remain visible on the landscape (1995); Heyerdahl et al. (1989) attempted to move a statue in and in satellite images (Lipo and Hunt, 2005). These roads provide a vertical position using only ropes and padding in a fashion in which direct information about some of the physical parameters neces- the statue was “wiggled” from side to side. While more realistic than sarily involved in statue movement. For example, a long stretch of horizontal arrangements, his attempt met with limited success moai road along the south coast has uphill gradients as steep as 13 given the statue’s unsuitable form, vertical instability, and great with an average of 2.9, and downhill slopes of 16 with an average friction quickly causing damage to its base. The most recent attempts of 2.8 (Fig. S1). Excavations of segments of the moai roads by Love have employed a wooden sledge with moai in horizontal (prone or (2001) also show that the roads are concave in cross section, supine) positions pulled over logs in a sliding motion (Van Tilburg a configuration that largely rules out use of transverse rollers. and Ralston, 2005). A conclusion favoring horizontal movement of Inline Supplementary Fig. S1 can be found online at http://dx. moai in prone or supine positions on wooden sledges slid over logs doi.org/10.1016/j.jas.2012.09.029. has seen some favor, appearing to fit the assumed impact the statue During field research on moai roads, we documented 62 statues cult played in the island’s deforestation (Diamond, 2005, 2007). located on and parallel to road features (Table S1). These moai can While experimental attempts highlight the plausibility of methods be explained as those in the process of being moved, but transport that could have been used to move statues, they have overlooked failed and they were abandoned. The explanation relies on several systematic variation in the statues and their broader context in the lines of evidence. First, the statues lay in a direction parallel to the archaeological record. road direction that runs away from the main quarry at Rano Raraku, C.P. Lipo et al. / Journal of Archaeological Science 40 (2013) 2859e2866 2861 an orientation consistent with transport. Second, these moai are not own. The reduction of the base width is thus related to changing the associated with platforms (ahu) where they would be erected at vertical orientation of the statue from a forward lean to a more their destinations. Third, like the statues remaining at the quarry, stable upright position (Fig. 2). none had eye sockets carved in them, in contrast to those erected Inline Supplementary Fig. S2 can be found online at http://dx. on platforms. Prehistoric islanders inserted into these eye sockets doi.org/10.1016/j.jas.2012.09.029. coral “eyes” with obsidian or red scoria pupils. As with the Finally, three additional observations of the road moai inform on crowning of red scoria hats (pukao) for many statues, adding the their transport. Thirty-seven percent of road moai are broken into eyes was a final step in completing moai with emplacement on two or more fragments consistent with breakage that resulted from platforms. None of the road statues in our analyzed set has these falling from a vertical position (e.g., Fig. S3). In addition, on the eye features, defining their incompleteness and status as in transit south coast road are two examples of moai that did not break and when abandoned. were partially buried at their base and can be explained by ancient The statues found along the roads have shapes that also workers who attempted to re-erect them by excavating a pit to distinguish them from those statues erected on ahu. The road moai restore them to an upright position to be ‘walked’ out on a ramp have statistically wider bases when measured relative to shoulder (e.g., Figs. S4 and S5). Second, 70% of road moai exhibit evidence of width than ahu moai (Fig. S2). Once statues arrived on platforms, fracture caused by force being applied in inverse vertical fashion prehistoric carvers modified the statues to decrease the width of along the lateral edges of the base. These fractures form broad the base relative to shoulder. In addition, while ahu moai stand in shallow concave scars from which a wide pressure flake was an upright fashion with their mass located well over their base, removed. Third, the positions of fallen statues are statistically non- road moai show a distinctive angled base that would cause the random (p < 0.05) with respect to the slope on which they lay statue to lean significantly forward, often more than 10. Indeed (Table S2). The majority of statues are found facedown when the such forward lean means that most could not stand upright on their road slopes downhill, and often on their backs when going uphill.

Fig. 2. Differences between road moai (in transit) and those modiﬁed moai found at ahu. Road statues have a distinctive forward lean that places the center of mass close to or slightly over the front edge. In addition, shoulder width relative to the width of the base is close to 1.0 giving the statue a lower center of mass. Ahu moai, in contrast, have a center of mass closer to the center of base. Shoulder widths for ahu moai tend to be wider than the base width, making the statue more top-heavy. 2862 C.P. Lipo et al. / Journal of Archaeological Science 40 (2013) 2859e2866

Inline Supplementary Figs. S3eS5 can be found online at http:// statues were lowered down into trenches from their quarry dx.doi.org/10.1016/j.jas.2012.09.029. bedrock sources above to the base of Rano Raraku where their These road and statue observations are explained by the backs were completed and other details carved, and then moved hypothesis that moai were transported in a standing position and upright out of these trenches remaining in a vertical position, some fell en route. Significantly, hypotheses for horizontal trans- including their final placement on ahu (Fig. 3). Otherwise, the port using wooden devices cannot account for these archaeological logistics of lowering and later raising tall multi-ton moai make little observations. Moreover, the field evidence clearly shows that sense and imply they remained vertical. Statues found along the

B C

F H

Fig. 3. Quarrying and positioning of moai. (A) On the upper slopes of Rano Raraku, an extinct volcanic vent, bedrock hyalotuff is carved into a moai. The figure is usually shaped from the top down leaving a narrow ‘keel’ connecting it to the bedrock. (B) To move the statue from its quarried location, carvers broke the keel and slid the moai downhill. Moai were then placed into standing positions in pits excavated near or at the base of the slope. Here, carvers finished removing material from the back of the moai and prepared statues for movement. (C) Statues were ‘walked’ out of the pit through excavated openings to moai roads. (D) Not all statues were removed from pits; many were left and following sedi- mentation only the heads and shoulders of these are now visible on the quarry slopes. Statues were moved along road features (E) that were prepared in advance. These roads often have a concave shape that matched the arc of the front edge of the moai base. The concavity of the roads provides constraints to keep the statue moving in the intended direction. Occasionally, the roads have stretches of “curbstones” (F) that line their edges. These stones may serve to keep sediment from filling in the concave roadbed. Along the roads are found examples of statues that failed in transport (G). These are often on the sides of roads demonstrating that the roads are realigned for statue transport as needed. Roads are kept relatively flat through the infilling of low areas and (H) and carving through hills and ridges. Once at their destination statues are walked up temporary ramps made of stone (I). The remains of these stone ramps were often used to form “wings” on the lateral edges of the platform. Upon reaching the top of the platform, the moai then are turned 180 (J) to face the ceremonial area that lay in front of the ahu. The moai are complete once they are reshaped to make them stand upright and their eye sockets are carved for coral insets. C.P. Lipo et al. / Journal of Archaeological Science 40 (2013) 2859e2866 2863 roads that fell during transport were often not re-erected given Inspection of the 3D digital model revealed that the moai would breakage or the difficulty of raising them. However, the question not stand upright on its own. For the moai replicated, like others arising from Pavel’s (1995) attempt at vertical transport remained that were in transit, the center of mass is positioned just over the unanswered: how would damage to the statue base from friction be point of the front edge and standing the statue on its base would minimized? If a short distance of vertical transport (in ‘wiggling’ or cause it to tip forward. Modification of the base for standing on an ‘shuffling’ as Pavel did) significantly damages the base, it seems ahu, therefore, would have been necessary had the statue reached improbable that the statues could have been moved several kilo- its destination, explaining the difference in the base to shoulder meters without causing catastrophic damage. ratios we document. The basal flaking and forward lean point to how the statue was fi 4. Evaluating the ‘walking’ moai ‘walked’ in an upright position without signi cant wear to the base caused by twisting on its basal surface as in Pavel’s earlier attempts. To address detailed problems of statue transport, we con- Rather than twisting in place, moai were tilted to the side and structed a three-dimensional model of an actual road moai (Fig. 4). allowed to fall forward in the direction of their lean. With a slight The original moai (12-220-01) is located along the ancient roadway tilt to the side, the statue then falls forward toward its inherent on the south coast of the island and is 7.35 m in height and 2.83 m front lean. As it falls, the moai rolls across its front edge and rotates in width. It had been transported ca. 3.54 km southwest from the slightly to the opposite side. In this way, friction between the base Rano Raraku quarry before it fell on its back intact on an uphill and the ground is minimized allowing for the conservation of fi slope of a south coast moai road. energy, increasing overall ef ciency and removing the potential for The 3D models of moai allowed us to measure the center of mass damage as the statue ‘walks.’ along each of three spatial dimensions. The center of mass was The difference between road moai and those that reach their ahu fi located in the middle of the statue in terms of its width as the destinations is signi cant, though previous experiments over- statues are generally left/right symmetric. The height of the center looked the systematic variations. Van Tilburg and Ralston (2005), of mass is also approximately in the center of the statue, midway for example, created a “statistically average moai” thereby masking between the base and the top of the head. But the center of mass in statue variability and modeling transport by confounding forms fi the depth dimension is remarkably forward relative to the base of designed to ‘walk’ versus those modi ed to stand erect on ahu fi the statue. This peculiar configuration cannot be explained by platforms. Signi cantly, a “statistically average moai” replica would a hypothesis of horizontal transport. In horizontal transport, one not be capable of transport by ‘walking’ since the vast majority of fi would expect to find the center of mass toward the base to facilitate statues were prehistorically modi ed to stand stably on platforms. raising the statue back up to its vertical orientation when placed on the ahu. 5. Demonstration of moai ‘walking’ The 3D model provides a means of evaluating how forces acting on it would result in motion. Like other road moai, this statue has To test the dynamics of motion and the practical constraints a distinctive forward lean and exhibits flaking on the lateral involved with walking a moai, we constructed a scaled replica of portions of the base. The plan view shape of the statue’s base is also a size sufficient to represent real-world challenges in moving these like other road moai with a dorsal edge is relatively straight and monumental figures. Despite previous experiments, ours is the first a ventral edge is broadly rounded. This rounded front edge is sha- use of a precise proportionally scaled replica of an actual road moai ped in such a way that it provides a continuous surface across shaped appropriately for transport, rather than standing erect on which the statue can roll as it is tilted from side to side. an ahu.

Fig. 4. 3D model of a road moai reconstructed using structure from motion algorithms in Microsoft Photosynth. The ﬁrst step involves taking uncontrolled overlapping photographs of the moai from as many points of view as possible. Second, the photos are synthesized to form a single representation of the moai surface using Microsoft Photosynth. The 3D models are constructed using point cloud data extracted from the completed photosynth. 2864 C.P. Lipo et al. / Journal of Archaeological Science 40 (2013) 2859e2866

perpendicular to the statue’s direction of travel. These lines were pulled in alternating fashion to ‘rock’ the statue from side to side. As the statue began to rock, it rolled forward along its front edge. Each roll caused the statue to take a ‘step.’ Supplementary video related to this article can be found at http://dx.doi.org/10.1016/j.jas.2012.09.029. Inline Supplementary Figs. S6eS10 can be found online at http://dx.doi.org/10.1016/j.jas.2012.09.029. When the center of mass is positioned in the center of the statue in the vertical direction, rocking the statue back and forth is relatively easy: the taller the moai, the greater the leverage enabling handlers to initiate its rocking. The width of the statue and the elongated head provides the statue the lateral stability and shape similar to a bowling pin. With such a shape, the statue we modeled can be tilted laterally as much as to 26 from vertical. The tilting to each side provides clearance on the opposite base edge for steps to be taken in an uphill fashion. By tilting the statue to one side, we Fig. 5. Scaled 3 m tall, 4.35 metric ton moai-replica ‘walking’ with teams handling three ropes at Kualoa Ranch, Hawai’i. Using a mold created from the 3D model, the could lift a lateral edge over 60 cm in height before the statue replica is formed of a density-corrected concrete and retains the same mass configu- would fall sideways. The degree to which the statue can be tilted to ration as the road moai on which it is based. the side provides the clearance necessary for the moai to climb uphill as well as provides the dynamic energy involved in each step. Downhill motion was also demonstrated, though we did not In four days of field trials in June and November of 2011 at explore the potential for moving the statue backwards while Kualoa Ranch, Hawai’i, a team of volunteers experimented with descending steep slopes (although no direct archaeological placement of ropes (100 hemp), cooperative tactics, and other evidence suggests this was attempted). logistical details (Fig. 5; Video S1; Figs. S6eS10). Initial vertical Describing the moai movement as ‘walking’ is conceptually placement of the replica using a crane confirmed that the moai consistent with our own pedal locomotion. Statue ‘walking’ and our could not stand on its own without inserts under the front edge own steps represent mechanics that can be modeled an inverse sufficient to tilt the center of mass back over its base. ‘Walking’ the pendulum: a simple pendulum that is turned upside down so the replica statue demonstrated that road moai can be moved effi- mass swings back and forth from a fixed base. Pendulums conserve ciently and with minimal friction and resulting wear (Fig. 6). The energy and can remain in motion for some duration as long as there statue’s shaped features are integrated into the range of motion is minimal friction during each swing. When moving to walk enabling ‘walking.’ Using the fewest inputs possible, we found that forward, one pivots on their foot placed on the ground. From that three ropes attached around the statue’s head were sufficient to pivot point, our center of massdlocated along the vertical center- initiate forward motion. We attached one rope near the top of the line of one’s bellydfollows the path of an arc as we lift our opposite head at the eyes of the moai and stretched it behind the direction of hip and swing a leg forward. One’s forward foot eventually hits the travel. This rope kept the statue position leaning slightly forward on ground and the arc slows to a stop in that direction. At that point its front edge, preventing it from falling too far forward. Tied to the kinetic energy is at a minimum, but potential energy on that side is same location at the eyes, we stretched two additional ropes at a maximum. As one falls forward into the next step, potential

Fig. 6. (A). Moai movement propelled by forward lean and side rocking using three ropes; solid lines with arrows show force of ropes; dashed lines show directions of movement. (B). Overhead view of moai walking motion. C.P. Lipo et al. / Journal of Archaeological Science 40 (2013) 2859e2866 2865 energy is converted back into kinetic energy, and continues move lean forward, but not fall to the ground. We were also able to easily forward. This is the basic physics of walking. ‘walk’ the statue uphill and downhill slopes as great as 6, turn and Moving large moai takes advantage of the same principle. change directions, as well as rotate the statue 180, requiring little, if During forward movement the statue tilts sequentially in opposite any, more space than its own base. While several near-falls were directions. The transfer of energy back and forth, however, does not averted by quick cooperative action of the rope handlers, we did involve the high friction “wiggle” motion as proposed by the Pavel drop the statue (face-forward) several times and re-positioned it (1995) method. Rather, road moai are designed so that the when the vertically using a large hydraulic crane. It would have been impos- statue rocks from one side to the other it also falls slightly forward sible to resurrect the statue by the handlers using ropes alone. The and rolls across its front edge. In this way, once the moai is set in great difficulty of raising a fallen statue may explain why so many motion it efficiently maintains the energy invested in the initial were abandoned along the roads even though they were not broken. tilting. Energy is smoothly converted from potential energy at the The ‘walking’ covered ground rapidly. In one continuous effort point of the statue’s greatest tilt to kinetic energy as it swings back we were able to move the statue about 100 m in just 40 min. In to the other side. Through this rocking and rolling motion, the contrast, Heyerdahl et al. (1989) estimated that a highly experi- statue takes incremental steps and moves forward. Once in motion, enced crew might “wiggle” a moai 100 m in a full day. The relatively relatively small amounts of energy must be added to the system rapid rate we achieved in ‘walking’ suggests that statues could have through gentle tugs on opposite sides. been moved several kilometers across the island in only a matter of This system of transportation is only possible, however, because weeks or months. It also follows that investment in statues was the statute is shaped to move in this fashion. Two factors are likely a part-time effort requiring relatively small groups of people. essential. First, the statues must lean forward so that they fall While large, our experimental moai replica is slightly smaller forward and any rocking motion also results in the statue rolling on than the mean size for statues across the island (ca. 4 m). Many the front edge. Second, the statue bases must be shaped in a way to examples of moai that were successfully moved are much larger, provide an edge across which the statue can roll. The outlines of the with the largest moved measuring about 10 m in height. Unlike bases for road moai have a flat back edge and distinct rounded front methods of movement using sleds, however, the ‘walking’ trans- edge (Fig. S11). Once moai reached their final locations, however, port method scales well as statues get larger. Larger statues are also prehistoric reshaping was necessary to allow them to stand upright taller, thus providing proportionally greater leverage for the rope in a stable fashion on their constructed stone platforms. Conse- handlers on the sides. At the same time, taller statues become quently, completed statues at the ahu could no longer be ‘walked.’ narrower overall and have increasingly longer and thinner heads Inline Supplementary Fig. S11 can be found online at http://dx. relative to their bodies (Figs. S12 and S13). This “bowling pin” shape doi.org/10.1016/j.jas.2012.09.029. not only gives moai their characteristic look, but it also means their ‘Walking’ the moai was achieved by creating inherent instability volume is decreased more than if they were scaled proportionally with its forward lean. In transit, statues would not be able to stand on in all dimensions and had a lower center of mass. Each of these their own, and without ropes or left stationary, they required stones, features reduces the effort that is required to tip the statue enough for example, wedged under them to balance the center of mass. This to initiate its rocking and thus ‘walking’ motion. observation seems to explain stone arrangements discovered in Inline Supplementary Figs. S12 and S13 can be found online at excavations at the bases of moai on roads (see Heyerdahl et al.,1989; http://dx.doi.org/10.1016/j.jas.2012.09.029. Richards et al., 2011). Thus, moai left standing along roads would Indeed, the larger the statues, the less likely they could have have been highly unstable and prone to fall without significant been moved in any way other than ‘walking.’ The fact that statues further modification of their form. get thinner with longer heads means that they become more fragile This observation rebuts early (Routledge, 1919) and recently and prone to breakage at the neck, as evident by frequent neck elaborated (Richards et al., 2011) speculations that road moai are breaks among the fallen road moai. Transporting large statues in not fallen attempts at moving a statue, but deliberate placements a horizontal fashion would put stress on the weakest point of the that mark a ritual path. The notion of a ritual path, however, bears statue. The hyalotuff that composes the statues is a relatively low- no impact on the issue of how the statues were transported, since it density material (Gioncada et al., 2011) that has good compressive has no direct empirical implications. Nor does such speculation strength, but low shear strength. Thus, while vertical objects can be explain why road statues have their characteristic shape and lack successfully made from the hyalotuff, horizontal shapes would be eye sockets as found in all ahu moai. Indeed, the presence of stones more prone to breakage. Even greater stresses would be placed on near the base of some of the road moai locations is easily explained moai if they also had to be levered up to a vertical position after as relating to the process of transport. Since few statues could have transport. In this way, the shapes of the road moai are clearly been moved from quarry to ahu in a single day, most would have inappropriate for transport methods other than ‘walking’ upright. had to have been left standing between moving efforts. With the instability inherent in the shape of statues and the relatively soft 6. Conclusions ground, this situation would have required stabilization around the base as has been noted in excavations. The successful transport of hundreds of multi-ton statues on In our experiments, remarkably small teams were easily able to prehistoric Easter Island has puzzled observers for centuries. initiate ‘walking’ the 4.35 metric ton road statue replica. Even with Modern research has focused on attempts such as hauling statues in our limited practical experience moving the statue, a minimum of horizontal positions on log contraptions with accompanying rollers only 18 people could achieve ‘walking.’ The lower limit for the or sliders. Significantly, such efforts have ignored systematic vari- number of individuals derives from the number needed on the sides ations in moai form. Statues were designed for movement and those to initiate rocking. The greatest initial energy input is the initiation of fallen on roads would have been further modified had they reached rocking of the statue from a static upright position. We achieved this their destinations at ahu. Moai patterns of breakage and wear, and with four individuals on each lateral rope. Once the statue began positions on roads relative to slope are also explained by a hypoth- rocking, however, relatively little input was required from each team esis of vertical ‘walking’ transport. As our research and experiments of lateral rope handlers as the statue motion conserved energy in its illustrate, these observations explain the archaeological record for repeated transitions from kinetic to potential to kinetic energy. Ten transport, and allowed us to deduce a falsifiable hypothesis for how people were needed to maintain the rear rope allowing the statue to the statues were moved. 2866 C.P. Lipo et al. / Journal of Archaeological Science 40 (2013) 2859e2866

In contrast to popular notions of sledges, rollers or sliders of Morrison, John O’Connor, Ted Ralston, Sergio Rapu, Jr., Timothy trees, the evidence shows that moai were specifically engineered to Rieth, Damion Sailors, Deborah Schechter, Sheela Sharma, Francisco ‘walk’ in an upright position achieved using only ropes, human Torres Hochstetter, Enrique Tucki and Janet Wilmshurst for their labor, and simple cleared pathways. A relatively small number of encouragement and generous support of our research. Finally, we people are capable of moving a statue; just 18 rope handlers could thank the many energetic, patient, and enthusiastic volunteers who ‘walk’ a statue weighing more than 4.35 metric tons. The statue helped us ‘walk’ a moai for the first time in centuries. evidence does not imply that a large population once existed on Easter, contrary to earlier and now popularized notions (e.g., Appendix A. Supplementary data Brown, 1924; Diamond, 2005, 2007). Apart from labor and engineering expertise, statue transport required only ropes; few if any Supplementary data related to this article can be found at http:// trees were required in statue transport. In fact, the primary vege- dx.doi.org/10.1016/j.jas.2012.09.029. tation on the island was the now extinct palm, Jubaea chilensis or a close relative, and given palm structure would likely not have References been suitable for use in building contraptions or making rollers that could support a great amount of weight. Material for ropes, Brown, J.M., 1924. The Riddle of the Pacific. T. Fisher Unwin, London. however, was abundant on the island as they were made from Diamond, J., 2005. Collapse: How Societies Choose to Fail or Succeed. Viking, New Triumfetta semitrioba York. a woody shrub ( that grows in disturbed Diamond, J., 2007. Easter Island revisited. Science 317, 1692. habitats, see Metraux, 1940). Consequently, statue making and Gioncada, G., Gonzalez-Ferran, O., Lezzerini, M., Mazzuoli, R., Bisson, M., Rapu, S., transport cannot be linked to deforestation, nor can forest clear- 2011. The volcanic rocks of Easter Island (Chile) and their use for moai sculp- ance for extensive cultivation of agricultural surplus to feed thou- tures. European Journal of Mineralogy 22, 855e867. Heyerdahl, T., 1989. Easter Island: the Mystery Solved. Souvenir Press, London. sands of statue workers, as some have supposed (see Diamond, Heyerdahl, T., Skjølsvold, A., Pavel, P., 1989. The ‘walking’ moai of Easter Island. 2005, 2007; Van Tilburg and Ralston, 2005: 299). The evidence Occasional Papers of the Kon-Tiki Museum 1, 55. for moai carving and transport points to activities by small-scale Hunt, T.L., Lipo, C.P., 2011. The Statues That Walked: Unraveling the Mystery of fi Easter Island. Free Press, New York. social groups rather than the product of laborers uni ed under Lipo, C.P., Hunt, T.L., 2005. Mapping prehistoric statue roads on Easter Island. a powerful centralized chiefdom. Here monumentality does not Antiquity 79, 309e317. imply large-scale social organization as assumed for many cases Love, C., 1990. How to make and move an Easter Island statue. In: Esen-Bauer, H.M. (Ed.), State and Perspectives of Scientific Research in Easter Island Culture. worldwide. Instead, we see moai carving and ‘walking’ as vivid Courier Forschungsinstitut Senckenburg, Frankfurt, pp. 139e140. expressions of costly signaling and evolutionary bet hedging in Love, C., 2000. More on moving Easter Island statues, with comments on the NOVA a competitive environment. Multiple lines of evidence, including program. Rapa Nui Journal 14, 115e118. Love, C., 2001. The Easter Island Moai Roads: an Excavation Project to Investigate the ingenious engineering to ‘walk’ statues, point to Easter Island as the Roads Along Which the Easter Islanders Moved Their Gigantic Ancestral a remarkable history of success in a most unlikely place. Statues. Unpublished Report on File at the P. Sebastian Englert Museum of Anthropology, Isla de Pascua, Chile. Metraux, A., 1940. Ethnology of Easter Island. Bishop Museum, Honolulu. Acknowledgments Pavel, P., 1995. Reconstruction of the transport of the Moai statues and Pukao hats. Rapa Nui Journal 9, 69e72. Funding for this project was provided by a grant from the Richards, C., Croucher, K., Paoa, T., Parish, T., Tucki, E., Welham, K., 2011. Road my National Geographic Expeditions Council. We thank Pacific body goes: re-creating ancestors from stone at the great moai quarry of Rano Raraku, Rapa Nui (Easter Island). World Archaeology 43, 191e210. Islanders in Communications and Nippon Television for their Routledge, K., 1919. The Mystery of Easter Island. Stifton, Praed & Co, London. additional support. We thank the National Park staff on Rapa Nui Thomson, W.J., 1889. Te Pito te Henua, or Easter Island, Report of the U.S. National fi for assistance in the field. We thank Hannah Bloch and Fernando Museum for the Year Ending June 30, 1889. U.S. Government Printing Of ce, Washington D. C, pp. 447e552. Baptista at National Geographic Magazine for their excellent work Torres Hochstetter, F., Rapu Haoa, S., Lipo, C.P., Hunt, T.L., 2011. A public database of on the July 2012 cover story and artwork. We also thank Maria archaeological resources on Easter Island (Rapa Nui) using Google Earth. Latin Awes, Andy Awes, the staff at Committee Films, Max Beach, John American Antiquity 22, 385e397. Van Tilburg, J., Ralston, T., 2005. Megaliths and Mariners: experimental archaeology Bredar, Robert DiNapoli, Brian Fagan, John Francis, Herman Ika, on Easter Island. In: Johnson, K. (Ed.), Onward and Upward! Papers in Honor of Marc Kelly, Jacce Mikulanec,John Morgan (Kualoa Ranch), Alex Clement W. Meighan, pp. 279e303.